Reputation: 196
I have a table named Reading_Hist containing columns such as Reading, Date, ID. This table contains history of the readings. example
+----+---------+-------------+
| ID | Reading | ReadingDate |
+----+---------+-------------+
| 1 | 12 | 9/12/2018 |
| 2 | 15 | 9/12/2018 |
| 1 | 16 | 9/5/2018 |
| 4 | 1 | 9/12/2018 |
| 3 | 65 | 9/12/2018 |
| 1 | 23 | 8/29/2018 |
| 3 | 25 | 9/5/2018 |
| 2 | 23 | 9/5/2018 |
| 4 | 3 | 9/5/2018 |
+----+---------+-------------+
I want to write a sql to display each ID and it's current Reading on first column, next reading taken a week before and third reading taken two weeks before and last trend of the reading.
Example Result below.
+----+---------+------+------+-------+
| ID | Current | Wk_1 | Wk_2 | Trend |
+----+---------+------+------+-------+
| 1 | 12 | 16 | 23 | Down |
| 2 | 15 | 23 | NULL | Down |
| 3 | 65 | 25 | NULL | UP |
| 4 | 1 | 3 | NULL | Down |
+----+---------+------+------+-------+
Upvotes: 0
Views: 54
Reputation: 37487
You can use aggregation to get the maximum day of readings per ID. Then left join the current readings, them of the last week and two weeks ago. Use CASE
to calculate the trend.
It could look something like:
SELECT x.id,
rh2.reading current,
rh3.reading wk_1,
rh4.reading wk_2,
CASE
WHEN rh2.reading > rh3.reading THEN
'Up'
WHEN rh2.reading < rh3.reading THEN
'Down'
WHEN rh2.reading = rh3.reading THEN
'-'
END trend
FROM (SELECT rh1.id,
max(rh1.reading_date) reading_date
FROM reading_hist rh1
GROUP BY rh1.id) x
LEFT JOIN reading_hist rh2
ON rh2.id = x.id
AND rh2.reading_date = x.reading_date
LEFT JOIN reading_hist rh3
ON rh3.id = x.id
AND rh3.reading_date = dateadd(day, -7, x.reading_date)
LEFT JOIN reading_hist rh4
ON rh4.id = x.id
AND rh4.reading_date = dateadd(day, -14, x.reading_date);
Of course this requires, that there are readings exactly 7 or 14 days from the last day of readings.
Upvotes: 1